Commit graph

77 commits

Author SHA1 Message Date
Bill Currie
0f30f0a133 [mathlib] Remove suspicious IS_NAN
The implementation looks wrong (more like infinity). Where it was used
is currently disabled, but the usages were replaced with C99's isnan.
2022-03-19 12:50:08 +09:00
Bill Currie
65af7fb4a4 [mathlib] Remove frustum global
It should never have been there and is now in the refdef (not its final
home: it should probably be part of the camera).
2022-03-19 12:33:12 +09:00
Bill Currie
16440bce2d [mathlib] Clean up AngleVectors comments a little
They're still slightly confusing, but the situation itself is confusing,
but the comments should be a little more helpful now as they are more
explicit about the orientation of the matrices and just which axis
points where.
2022-03-14 11:51:50 +09:00
Bill Currie
7c07118541 [mathlib] Clean up AngleVectors comment
I finally spent the time to work out what it was trying to say. Still
not sure it's clear, but what is clear is that there was probably some
disagreement at Id about the orientation of the world.
2022-03-01 14:52:45 +09:00
Bill Currie
9d819254d4 [util] Make a number of improvements to SEB
Attempting to vis ad_tears drags a few lurking bugs out of
SmallestEnclosingBall_vf: poor calculation of 2-point affine space, poor
handling of duplicate points and dropped support points, poor
calculation of the new center (related to duplicate points), and
insufficient iterations for large point sets. qfvis (modified for
cluster spheres) now loads ad_tears.
2021-07-30 14:57:47 +09:00
Bill Currie
6d312aaa63 [simd] Check the distance to the affine point
As per usual, fp math finds a way to confound any epsilon test. So
rather than relying entirely on test_support_points, check the distance
from the sphere center to the affine point and break out of the loop if
the distance is small enough (< 1% of the current radius). This allows
qfvis to load ad_tears without hacks.
2021-07-29 15:15:14 +09:00
Bill Currie
45aa8e6504 [util] Loosen affine test epsilon for SEB
Scaling the checks by 1e-6 was a little too tight for very small
triangles, but 1e-5 seems to work well. This fixes SEB getting stuck for
a ridiculously small (for quake) triangle in ad_tears (probably resulted
from some bad math in qfbsp when generating the portal file from the
bsp).
2021-07-29 15:03:54 +09:00
Bill Currie
ef9b04ba83 [util] Get tests working with sse2
It seems that i686 code generation is all over the place reguarding sse2
vs fp, with the resulting differences in carried precision. I'm not sure
I'm happy with the situation, but at least it's being tested to a
certain extent. Not sure if this broke basic (no sse) i686 tests.
2021-06-01 18:53:53 +09:00
Bill Currie
0167456b21 [util] Loosen up SEB relative epsilon a little
1e-8 is a little too tight, the normal was 4.5x the computed epsilon.
Going to 1e-6 still keeps things tight.
2021-03-27 15:04:13 +09:00
Bill Currie
693225a16f [util] Add a fixme for a comment
The code itself is fine, but the comment is rubbish because it's
confusing, though essentially correct.
2021-03-10 18:05:12 +09:00
Bill Currie
62f3e1f428 [util] Calculate quaternion to rotate between two vectors
The calculation fails (produces NaN) if the vectors are anti-parallel,
but works for all other combinations. I came up with this implementation
when I discovered Unity's Quaternion.FromToRotation could did not work
with very small angles. This implementation will produce a usable
quaternion below 0.00255 degrees (though it will be slightly larger than
unit). Unity's failed such that I could see KSP's skybox snap while it
rotated around my test vessel.
2020-12-21 14:14:10 +09:00
Bill Currie
e991c44232 [util] Make a minor improvement to QuatMultVec
Switch from using addition to multiplication for doubling the value.

Also, fix lib/util/test Makemodule to actually work (too much c&p
coding).
2020-12-21 14:13:57 +09:00
Bill Currie
86b5b30b45 Merge branch 'master' into vulkan 2020-06-25 14:03:52 +09:00
Bill Currie
7ae047654b [util] Rework SEB affine/convex testing
When I ported SEB to python, I discovered that I apparently didn't
really understand the paper's description of the end condition and the
usage of the affine and convex sets for center testing. This cleans up
the test and makes SEB more correct for the cases that have less than 4
supporting points (especially when there are less than 4 points total).
2020-06-21 17:07:54 +09:00
Bill Currie
e164002050 Make QuatToMat faster and more accurate
The better accuracy is for specific cases (90 degree rotations around a
main axis: the matrix element for that axis is now 1 instead of
0.99999994). The speedup comes from doing fewer additions (multiply
seems to be faster than add for fp, at least in this situation).
2019-07-23 08:52:15 +09:00
Bill Currie
c727f6a130 Rewrite QuatMultVec to be faster
Could be faster still using SSE, but that's another project.
2019-07-06 14:45:40 +09:00
Bill Currie
f58c2fef5a Switch QF's quaterions from wxyz to xyzw.
After messing with SIMD stuff for a little, I think I now understand why
the industry went with xyzw instead of the mathematical wxyz. Anyway, this
will make for less pain in the future (assuming I got everything).
2018-10-12 14:33:55 +09:00
Bill Currie
6eec76dd49 Implement Fischer's SEB algorithm (for 3d).
Now we can get tight (<1e-6 * radius_squared error) bounding spheres. More
importantly (for qfvis, anyway) very quickly: 1.7Mspheres/second for a 5
point cloud on my 2.33GHz Core 2 :)
2013-03-13 17:10:55 +09:00
Bill Currie
b6d4766201 Add a function to find the circumsphere of up to 4 points.
It seems to be a little sloppy (an error of a bit less than 1e-4). This
might be why I'm having trouble with my SEB code.
2013-03-13 14:35:30 +09:00
Bill Currie
0cd6d93030 Add a function to get the barycentric coords of a point.
It "works" for lines, triangles and tetrahedrons. For lines and triangles,
it gives the barycentric coordinates of the perpendicular projection of the
point onto to features. Only tetrahedrons are guaranteed to reproduce the
original point.
2013-03-12 14:16:42 +09:00
Bill Currie
6f484ee757 Add support and tests for 3x3 matrices. 2012-08-18 16:29:57 +09:00
Bill Currie
b5add680f7 Add Mat4as3MultVec.
to transform a vertex without the translation (eg, for normals) without
having to have separate matrices.
2012-05-20 15:20:18 +09:00
Bill Currie
4ba69b511b Make quat_origin the right type.
It's quat_t, not vec3_t :P.
2012-05-15 21:08:46 +09:00
Bill Currie
9f253454e4 Implement and test Mat4Inverse(). 2012-05-15 21:08:44 +09:00
Bill Currie
fa6270322f Make Mat4Decompose a little more intuitive.
Params are now in application order (good thing not much code uses it yet).
2012-05-15 21:08:44 +09:00
Bill Currie
de131c18ad Add VectorShear and really test Mat4Decompose.
And the tests really exercised VectorShear (first attempt had things
messed up when more than one shear value was non-zero). Also,
Mat4Decompose wasn't orthogonalizing the z axis row. Oops. Anyway,
Mat4Decompose is now known to work well, and the usage of its output is
understood :)
2012-05-15 21:08:44 +09:00
Bill Currie
8cea6b0a1e Fix a scaling snafu in Mat4Decompose.
Found when checking how rotation affects non-uniform scaling (haven't
written proper tests yet: need to figure out how).
2012-05-15 21:08:44 +09:00
Bill Currie
82abd5e426 More matrix tests. 2012-05-15 21:08:44 +09:00
Bill Currie
730f9668bd Add support for mat4 * vec3 2012-05-15 21:08:44 +09:00
Bill Currie
50af8a562e Add tests for matrix decomposition and fix the bugs. 2012-05-15 21:08:44 +09:00
Bill Currie
215d2e9747 Add a function to decompose a matrix.
The function gives rotation(quaternion), scale, shear, and translation (all
three as vectors), shear is (XY, XZ, YZ).
2012-05-15 21:08:44 +09:00
Bill Currie
8791b35e55 Add support for half floats.
iqm and OpenGL use them, so they might come in handy. The tests use values
from wikipedia and a couple extra.
2012-04-26 20:55:11 +09:00
Bill Currie
efaef89c5f Add a shortcut function for quaterion/vector multiplication.
I got the idea from blender when I discovered by accident that quat * vect
produces the same result as quat * qvect * quat* and looked up the code to
check what was going on. While matrix/vector multiplication still beats the
pants off quaternion/vector multiplication, QuatMultVec is a slight
optimization over quat * qvect * quat* (17+,24* vs 24+,32*, plus no need to
to generate quat*).
2012-04-26 09:48:08 +09:00
Bill Currie
bc1b483525 Nuke the rcsid stuff.
It's pretty useless in git.
2012-04-22 10:56:32 +09:00
Bill Currie
cbd13ac54e Implement QuatExp and fix up QuatBlend's param names. 2012-01-12 21:22:06 +09:00
Bill Currie
ea2ae1436c Fix some bugs in the new matrix functions/macros. 2011-12-31 13:20:31 +09:00
Bill Currie
35bf957d08 Add a 4x4 matrix type and some functions for it. 2011-12-30 20:11:53 +09:00
Bill Currie
91e65b6c80 Rename mplane_t to plane_t and clean up the mess.
I got rather tired of there being multiple definitions of mostly compatible
plane types (and I need a common type anyway). dplane_t still exists for
now because I want to be careful when messing with the actual bsp format.
2011-11-28 20:54:34 +09:00
Bill Currie
12e238f806 Add a function to convert quake angles to a quaternion.
The test works by comparing the result of AngleVectors with forward,
right, and up vectors rotated by the quaternion resulting from AngleQuat.
2011-09-04 09:58:12 +09:00
Bill Currie
fe3cdbcf5a fix a snafu in quattomat 2011-09-04 09:58:12 +09:00
Bill Currie
eb205d73fd Add a function to convert a quaternion to a rotation matrix. 2011-09-04 09:58:12 +09:00
Bill Currie
33db650161 Add QuatInverse to calculate 1/q.
There's still no direct quaternion division because I don't know if the
definition of q1/q2 is (1/q2)*q1 or q1*(1/q2) (it does make a difference).
2011-01-18 22:22:18 +09:00
Bill Currie
13503198e5 Add quat_origin convenience constant.
Same as vec3_origin in that all elements are zero.
2011-01-11 08:41:40 +09:00
Bill Currie
99c0954b47 the big dso visibility patch :). Sure, we have to have unique names for static builds, but with controlled visibitly we should get faster program loads (although this isn't C++, so it's not as bad) and complex plugins are cleaner. 2007-03-10 12:00:59 +00:00
Bill Currie
3fb03fc2be hah, should have been using __attribute__((used)) all that time (rather
than __attribute__((unused))). fixes the missing console in -x11
2005-08-04 15:27:09 +00:00
Bill Currie
4c4f40f137 fix a few bugs. basics seem to be working 2004-04-08 05:13:24 +00:00
Bill Currie
39685d0c63 more readability and add the quaternion functions 2004-04-07 18:01:45 +00:00
Bill Currie
19ef80c555 this should fix static win32 builds 2004-01-21 00:15:35 +00:00
Bill Currie
7ead5a91f8 don't let gcc automaticly inline functions but make gcc inline bigger
functions when told to. also make gcc warn if it can't inline a function.

Explicitly inline several functions (including moving VectorNormalize to
mathlib.h so it /can/ be) resulting in a 5.5% speedup for spam2 (88 to 92
fps)
2003-08-11 06:05:07 +00:00
Bill Currie
6a94607d9d rename VectorMA to VectorMultAdd, add VectorMultSub and use it in a few
places. Not certain how much speed difference it makes :/
2003-08-08 15:25:53 +00:00